[CP2K-user] [CP2K:19792] Cons. Qty for different ensembles + Pressure fluctuation

[Léon Luntadila Lufungula]

Hi Jurg and Matthias,

Jurg:
1) Okay, good to know that I can reduce the fluctuations by increasing the 
system size. Still, are fluctuations in the order of kbar not too large 
when the desired pressure is 1 bar?
2) I'm using the CSVR thermostat with a timecon of 100 fs and the barostat 
is the standard one with a timecon of 100 fs by just including the 
following lines to the input and changing the ensemble to NPT_I: 
&BAROSTAT
  TIMECON 100
&END BAROSTAT
2-continued) I do see now that the BAROSTAT section also has a THERMOSTAT 
subsection. Should the thermostat be defined here in the case of an NPT 
calculation or is it in the THERMOSTAT subsection of MD as is the case for 
an NVT run? If not, what is the use of the BAROSTAT/THERMOSTAT subsection?

Matthias:
Thanks for pointing this out to me! I remembered noticing that the value 
for GMAX was too small this weekend, but I forgot to adjust it accordingly. 
I'll try to re-run the calculations with the inputs in the regtest-water 
directory to see if I get better results! Maybe one more question if that's 
okay for you. I'm also running / getting ready to run AIMD simulations and 
these are quite heavy so it takes a long time to simulate even a couple of 
ps. If I see a drift in the constant of motion like in the pictures above, 
is there a way of knowing if this is due to an error in the input or just 
because the system has not equilibrated yet? Or is this just a matter of 
simulating for longer and longer times to see if it goes towards an 
equilibrium or keeps drifting? Perhaps the "Energy drift per atom [K]" in 
the output can be of some help, but I'm not certain how it's calculated or 
how I can use it to check my trajectory...

Thanks for all the help!

Kind regards,
Léon
On Monday 22 January 2024 at 14:29:09 UTC+1 Krack Matthias wrote:

> Hi
>
>  
>
> The GMAX parameter for Ewald is a bit sloppy in these inputs. Suggested 
> values are 1 to 2 points per Angstrom based on the cell edge lengths, i.e. 
> GMAX 20 is numerically more appropriate. In the example I gave you the GMAX 
> value is automatically adapted to the cell size.
>
>  
>
> Best
>
>  
>
> Matthias
>
>  
>
> *From: *cp... at googlegroups.com <cp... at googlegroups.com> on behalf of Léon 
> Luntadila Lufungula <Leon.luntad... at uantwerpen.be>
> *Date: *Monday, 22 January 2024 at 14:13
> *To: *cp2k <cp... at googlegroups.com>
> *Subject: *Re: [CP2K:19788] Cons. Qty for different ensembles + Pressure 
> fluctuation
>
> P.S. These calculations were done based of the input files in 
> https://www.cp2k.org/exercises:2018_uzh_acpc2:mol_sol, because the Fist 
> regtest-water folder was not yet available at the time, but the FF params 
> are correct, the coordinates are just different than the one molecule box 
> in the tests directory.
>
>  
>
> On Monday 22 January 2024 at 14:03:06 UTC+1 Léon Luntadila Lufungula wrote:
>
> Dear Matthias,
>
>  
>
> I have similar results for a box of 256 water molecules:
>
> NVT:
>
>
> NPT: (fluctuations are a bit smaller but still in the range of several 
> kbar...)
>
>  
>
> You say that the sample should be equilibrated well in advance before 
> starting the sampling run, so do you mean I should run an NVT simulation 
> for longer until it equilibrates before running an NPT from the thermalized 
> end structure? Or do you mean both NVT and NPT should be run for longer 
> until the Cons. Qty. is also equilibrated? Because in both cases the 
> temperature seems to be equilibrated quite well and the pressure in the NPT 
> ensemble is also constant albeit with very large oscillations...
>
>  
>
> Kind regards,
>
> Léon
>
>  
>
>  
>
> On Monday 22 January 2024 at 13:40:05 UTC+1 Krack Matthias wrote:
>
> Dear Léon
>
>  
>
> a 32 H2O cell is still a relatively small system which can cause large 
> fluctuations. I suggest to try at least the 6x6x6 = 216 H2O input 
> <https://github.com/cp2k/cp2k/blob/master/tests/Fist/regtest-water/H2O-TIP3P-6x6x6.inp>. 
> The sample should be equilibrated well in advance before starting the 
> sampling run.
>
>  
>
> HTH
>
>  
>
> Matthias
>
>  
>
> *From: *cp... at googlegroups.com <cp... at googlegroups.com> on behalf of Léon 
> Luntadila Lufungula <Leon.luntad... at uantwerpen.be>
> *Date: *Monday, 22 January 2024 at 12:56
> *To: *cp2k <cp... at googlegroups.com>
> *Subject: *[CP2K:19784] Cons. Qty for different ensembles + Pressure 
> fluctuation
>
> Dear all,
>
>  
>
> I ran several simulations on a box of 32 water molecules with the TIP3P 
> force field:
>
> 1) 1 ns NVE trajectory with timestep 0.5 fs
>
> 2) 1 ns NVT trajectory at 298.15K with timestep 0.5 fs and thermostat time 
> constant of 100fs
>
>  3) 1 ns NPT_I trajectory restarted from the final structure of the NVT 
> trajectory at 298.15K and 1 bar with timestep 0.5 and thermostat and 
> barostat time constants of 100 fs
>
>  
>
> I plotted the Cons. Qty [a.u.] vs. Time [fs] for the three simulations and 
> saw that the constant of motion was only more or less constant for the NVE 
> trajectory. The NVT and NPT trajectories both had strong oscillations and 
> significant drift of the constant of motion (see below) and I was wondering 
> if this is normal behaviour or that this means that the system is not yet 
> equilibrated? I would think that the Cons. Qty. should remain constant but 
> I can also understand that the thermostat and barostat might influence its 
> behaviour in a way.
>
> NVE:
>
> NVT:
>
> NPT:
>
> The simulations of the water box were small test cases, but I also need to 
> know what behaviour to expect when going to my production runs with my 
> systems of interest...
>
>  
>
> Additionally, in the NPT ensemble, the pressure seems to fluctuate very 
> strongly from -20000 to +20000 bar which seems to be quite excessive to me. 
> Is this normal? I've also seen this asked in another post 
> <https://groups.google.com/g/cp2k/c/gxA8r6wmyF4/m/k8QT7iglAAAJ>, but that 
> person did not get a reply unfortunately....
>
>  
>
> Thanks in advance for the help!
>
>  
>
> Kind regards,
>
> Léon
>
> -- 
> You received this message because you are subscribed to the Google Groups 
> "cp2k" group.
> To unsubscribe from this group and stop receiving emails from it, send an 
> email to cp2k+uns... at googlegroups.com.
> To view this discussion on the web visit 
> https://groups.google.com/d/msgid/cp2k/76d3e94b-4a67-4fcc-8861-78a566ebb107n%40googlegroups.com 
> <https://groups.google.com/d/msgid/cp2k/76d3e94b-4a67-4fcc-8861-78a566ebb107n%40googlegroups.com?utm_medium=email&utm_source=footer>
> .
>
> -- 
> You received this message because you are subscribed to the Google Groups 
> "cp2k" group.
> To unsubscribe from this group and stop receiving emails from it, send an 
> email to cp2k+uns... at googlegroups.com.
>
> To view this discussion on the web visit 
> https://groups.google.com/d/msgid/cp2k/e6d4002e-91a7-45cd-817c-adc66616d1d6n%40googlegroups.com 
> <https://groups.google.com/d/msgid/cp2k/e6d4002e-91a7-45cd-817c-adc66616d1d6n%40googlegroups.com?utm_medium=email&utm_source=footer>
> .
>

-- 
You received this message because you are subscribed to the Google Groups "cp2k" group.
To unsubscribe from this group and stop receiving emails from it, send an email to cp2k+unsubscribe at googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/cp2k/9ee8f63c-639e-4708-8a49-19f2cd8b87e4n%40googlegroups.com.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20240122/39770e60/attachment-0001.htm>